finding sequence size

sam@ · 12-24-2013, 01:53 AM

Hi All,

I have a file which has sequences which look like this

>String1
aqwertrtrytytyuuijhgddfghhhghhgjhjhhsswekrkmygppdslxmvbnhkwqalldrtjbllnlnlnnnvc
>String2
qwwerrtyuiopasdfghjmnbvfklzxerbvcwghjjkoowwqerrtggbddqsdfgaqwcxzakjtyugfsdefrtgyhujiknbbbbcdcdcxsxsx zxzxcvcfcdcg
>String3
rtyhujrfedwsqavfbgnhmjklopoiuytiuytrewqxszavfbgnhmjkjgdaaarftgwqwqsxddfcazxshjklopute
...

>String120..

so on

Whenever I wanted to find the size of any particular sequence (say string 1) I used to delete the remaining part of the file( ie from string2 till end) ,and find file size using du -h .This gave me the size of string1.

However is there an easier way that I could find say size of (string1 or say string2 )rather than deleting the remaining part to make a new file.

druuna · 12-24-2013, 03:49 AM

@sam@: Can you be more specific about what it is you are after?

You talk about the size of a string and using du to determine this. du reports the size of a file, normally reported in the amount of blocks needed for disk storage.

Using your string1 as an example, du will show:

Code:

$ du test.file
4       test.file

This tells you that 4.0k is needed to store this file on disk. This reported size will stay the same if you add a few characters to the string (up to a point, after which the size will become 8.0k)

The actual file itself isn't 4.0k:

Code:

$ du -b test.file
80      test.file

# or using stat:
$ stat test.file 
  File: `test.file'
  Size: 80              Blocks: 8          IO Block: 4096   regular file

80 bytes are needed.

The length of the string itself is 79 characters("bytes") the extra byte is the carriage return, which is part of the file and not the string itself.

So, what is it you are actually trying to determine?

jpollard · 12-24-2013, 04:34 AM

You can also always use awk to determine the line length...
or perl...

sam@ · 12-24-2013, 06:25 AM

Basically What I meant to say is whenever I wanted to find the length of any string (take case of finding length of string1)
i used to create a new file having

>String1
aqwertrtrytytyuuijhgddfghhhghhgjhjhhsswekrkmygppdslxmvbnhkwqalldrtjbllnlnlnnnvc

and delete the remaining part of original file.
Then using a ls -la i used to find the size of this newly created file.

My query is there a way i could find the file length of string1 string 2 without having to split the file

sam@ · 12-24-2013, 06:28 AM

Intuitively if I need to find the size of string1 ,it will calculate from >string1 till >string2 and give me the bytes(file size) between them.

For finding size of string2, it will start from >string2 to>string3 and find the bytes between them and so on.

druuna · 12-24-2013, 06:52 AM

You are still mixing length, bytes and (file) size...

Ok, have a look at this:

Code:

$ cat  test.file
>String1
aqwertrtrytytyuuijhgddfghhhghhgjhjhhsswekrkmygppdslxmvbnhkwqalldrtjbllnlnlnnnvc
>String2
qwwerrtyuiopasdfghjmnbvfklzxerbvcwghjjkoowwqerrtggbddqsdfgaqwcxzakjtyugfsdefrtgyhujiknbbbbcdcdcxsxsx zxzxcvcfcdcg
>String3
rtyhujrfedwsqavfbgnhmjklopoiuytiuytrewqxszavfbgnhmjkjgdaaarftgwqwqsxddfcazxshjklopute

With that in mind:

Code:

# get length of String2
$ thisString=$(sed -n '/String2/{n;p}' test.file)
$ echo "${#thisString}"
113

# get length of String1
$ thisString=$(sed -n '/String1/{n;p}' test.file)
$ echo "${#thisString}"
79

# get length of String3
$ thisString=$(sed -n '/String3/{n;p}' test.file)
$ echo "${#thisString}"
85

The sed command looks for StringX and print the next line (the actual string), which is stored in a variable named thisString. The echo "${#thisString}" prints the length of the found string.

If this is not what you are after then please provide more details (also provide an input and expected output example).

EDIT: I just noticed that there is a space in string number 2, which makes it 2 strings. The above sed solution ignores this (the space is counted as 1 character).

jpollard · 12-24-2013, 06:53 AM

Read the bash manpage.

You will find the answer there. If the string is in a shell variable s, then ${#s} will be its length.